Dr Janet Fanslow at University of Auckland’s medical school has again been given extensive publicity for ongoing findings in her study on violence against women in NZ, e.g.
1. “Ten per cent of pregnant women victims of violence” NZ Herald 13/08/2008
2. “We all need to be aware of how to protect our country’s young” NZ Herald 29/10/2007
3. “Violence linked to abortion rate” NZ Herald 14/08/2008
Voices of reason on MENZ have frequently challenged exaggerated claims made by feminist groups and the domestic violence industry concerning violence against women in NZ and other western countries. Those challenges pointed to high-quality research studies showing much more moderate findings about violence towards women and that might have bothered also to measure violence towards men. In society generally, men are much more often than women the victims of violence and the most serious violence. And that’s not taking into account the huge gender inequality concerning men killed or disabled in wars, civil emergencies and the work roles expected of them.
So how are feminist groups allowed to make continued claims like ‘1 in 3 women will face violence from their partners’? Those groups will point to their preferred research in support of their claims, and a closer look at their methodology, definitions and questions asked will show how they managed to obtain their figures. The ‘1 in 3’ figure appears to have come from Dr Fanslow’s research (See “One in three women assaulted by partner”, NZ Herald 26/11/2004.)
Dr Fanslow’s most recent claimed finding was that nearly 40% of women who asked for help after physical or sexual violence did not get help. However, it turns out that those women had ‘told somebody’ about such claimed violence but had not necessarily asked for help and/or had not asked someone in a position to help, so the ‘neglected victim’ story is already thrown in doubt. They may specfically have asked their confidant not to take the matter further, as most of us will have done on occasion for various reasons.
Dr Fanslow’s research was based on World Health Organization (WHO) research that continues to provide one major source of ongoing support around the world for what may be exaggerated claims (and what are definitely gender-unbalanced claims), about violence. Although Dr Fanslow’s previous research papers gave somewhat different definitions of ‘violence’ to those used in the original WHO study, it was not clear to what extent her definitions were reflected in the WHO interview methodology that her replication study used. I took a close look at the WHO research and found much to explain why it has been such a convenient tool for extending feminist propaganda.
The ‘WHO Multi-Country Study on Women’s Health and Domestic Violence’ was initiated in 1997, and by 2003 researchers had interviewed 24,000 women in 10 countries. Since then the study’s methodology including questionnaires etc has been replicated in other countries including NZ (Fanslow et al, 2004). I could find no critical analysis of the WHO methodology anywhere in the literature. This is surprising given obvious scientific and logical weakness inherent in the studies. Perhaps the relevant scientific community has been cowed by the long list of important-sounding advisors to the WHO study, or by the political power now enjoyed by feminist lobbyists both in governments and within professional groups. The following comments apply to the original WHO methodology as published, but replications of WHO’s study can be expected to suffer from similar scientific weaknesses.
One basic problem is that the measurement is of what people say to an interviewer, not actual events. We have no way of knowing the correlation between what is reported and what actually happened. No attempt was made in the initial WHO studies to check whether the verbal claims of the interviewed women were supported by any medical or police records. While that may be excusable due to poor public records in some of the 3rd-world countries initially targeted, when the study is replicated in developed countries like New Zealand there seems no good reason (except dishonesty) to omit any attempt at establishing external validity. Given prevailing beliefs that women continue to be powerless, discriminated against, abused and repressed more than men are (which may or may not be true depending on your country), many respondents in such interviews may be inclined to make up, exaggerate or to reinterpret the kinds of experiences that the interviewer is clearly looking for in the hope that the resulting published report will advance women’s cause.
Another fundamental threat to the scientific integrity of the WHO-based studies is that they are heavily based on a belief that the women respondents will tend not to disclose abuse and therefore needed various forms of encouragement to do so. For example, the women were asked about sexual abuse, then even if they did not disclose any they were asked specifically whether any from a list of various potential abusers had sexually abused them (e.g. father, uncles, teachers). Then later in the interview they were again asked and instead of being required to say anything they were given two pictures from which to choose their answer, one of a woman crying for ‘yes’ and one of a woman smiling for ‘no’. Apart from ambiguity introduced by such pictures, and apart from the value-laden messages about victimization of women given by the pictures and throughout the interviews in many other ways, the repeated attempts to get women to agree they had been sexually abused seemed little more than extra shots to try for the bulls-eye, and will have introduced very strong demand characteristics into the data. The respondents experienced their initial honest answers as being ignored and undesired, they understood what the interviewers wanted them to say and many of them eventually obliged.
Further, the interviewers (all female) were trained to extract the ‘yes’ responses they were looking for. That this introduced bias into the data was supported by the finding that when ‘professional interviewers’ were employed to meet a shortfall in one country, and those interviewers received only one day’s ‘training’ compared with several days for the study’s normal interviewers, those professional interviewers obtained considerably fewer disclosures of violence than did the study’s own trained interviewers. Of course, the researchers interpreted this as showing that the full ‘training’ was necessary to help women disclose their difficult experiences. The researchers seemed not to consider for a moment that any of the responses they extracted might be fallacious or exaggerated in response to interviewer demand characteristics. The researchers assumed that ‘no’ responses to violent experiences were likely to be lies, but if after repeated questioning a ‘yes’ response was extracted this was immediately assumed to be the truth.
The ‘structured interview’ method introduced further threats to data validity. Instead of simply presenting a questionnaire (that has established reliability and validity characteristics) or even asking standard questions and recording the answers, the interviewers were allowed to expand upon the questions and to ask additional non-standard questions to ‘help’ respondents arrive at their answers. This was related to a priority given in the research to show care, empathy and help for respondents who might be remembering traumatic experiences. Fair enough, but where was the care and respect to protect the accuracy of the resulting data? It has been well established that memory is fallible and easily distorted, and both men and women can easily be convinced to see themselves as victims. But no consideration was apparent in the WHO methodology to reduce the risk of false positives; on the contrary, the design appeared deliberately aimed at maximizing positive responses regardless of validity.
I note that a performance indicator used by supervisors in evaluating the interviewers was ‘the number of disclosures of violence obtained’. This was ridiculous and obviously will have compromised the validity of the data.
Another problem (seen also in most feminist advocacy research) is one of defining ‘violence’ to include all manner of events that any reasonable person would not view as violent. The questions asked about physical and sexual violence included overly general and subjective ones that almost certainly inflated the abuse data. For example, physical violence included being pushed or shoved. Few of us male or female will never have been pushed or shoved, and answering yes to this question could mean anything from an accidental attempt by both people to enter a door at the same time, to a quite gentle pushing away when a partner is shouting in your face, to an intentionally violent push intended to cause injury. Which of these really shows violence? Obviously for the researchers they will all do nicely to bolster the sensationalist figures they wanted.
For sexual abuse the respondents were asked various questions (in addition to the two-picture method used later in interviews) including ‘respondents were asked how old they were at their first sexual experience (Question 1004)’, and whether it had been ‘something they had wanted to happen, something they had not wanted but that had happened anyway, or something that they had been forced into (Question 1005)’. The answer yes to this wide range of options may or may not mean sexual abuse happened. Whether someone wanted their first sexual experience to happen says nothing about their actual participation or consent at the time. We might get carried away by passion, perhaps disinhibited by alcohol, then later regret what we did and think we didn’t really want it to happen, but that does not mean we were abused. Not really wanting sex but ‘it happened anyway’ may well describe the subjective experience of many males and females, given cultural guilt about sexuality and fear of consequences if caught by parents etc, but it doesn’t necessarily mean anyone abused us. Nevertheless, the researchers happily included all yes answers to show violence against women.
WHO published more detailed figures about responses to particular areas of questioning. Here, a tendency to obfuscate continued through the strange categorization of the data. Every statistic likely to reflect (claimed memory of) true violent abuse was thrown together with responses to questions that may well reflect events that did not involve real abuse. For example, the data purporting to reflect (claims of) violent victimization in the previous 12 months were grouped into ‘slapped or threw something’, ‘pushed or shoved’, ‘hit with a fist or something else’, ‘kicked or dragged’, ‘choked or burnt’, ‘threatened or had weapon used against them’. It is simply not possible from these groupings to derive reliable estimates of (claimed) true violent victimization. For example, most would agree that being ‘slapped’ would generally be true violence, with some exceptions e.g. an affectionate slap on the arm in response to a lewd joke. However, combining being slapped with having something thrown at one simply cast doubt on the extent to which the responses reflected (claims of) true violence at all. Things may be thrown at us non-violently, e.g. as a form of passing something across a space, or as part of a game. And light, harmless things if thrown even in anger can hardly amount to violence; e.g. “well read this letter for yourself then”. Yet under the demand characteristics of these interviews a respondent’s vague memory that something was thrown at her allows her to answer ‘yes’ and this counts towards yet another violent act. Similarly, being ‘kicked’ will generally amount to ‘being subjected to violence’ but the memory of being ‘dragged’ could relate to all manner of incidents from true violence to gentle encouragement to leave a party when one is drunk and it’s late, actions that both men and women will do and that only extreme ideologues would ever view as violence. Similarly again, ‘threatened or had weapon used against them’ lumps together experiences of serious violence with that of various interpreted situations that may never have involved truly violent threats (e.g. someone leaning on their axe while arguing with you, someone gesticulating with a walking stick to emphasize the points they are making, someone referring to a threat posed by a third party: “the police will attack you with batons if you behave that way in public” someone making an off-hand remark that would never be taken seriously: “You should be shot for saying such a thing” said by someone who has never owned a gun. Why have these responses been combined in these particular ways? Was it to inflate each response category of valid, serious violence by including responses to questions that may or may not reflect true claimed violence?
From what I have seen the replications using the WHO methodology made no attempt to measure violence against males either child or adult so could not reasonably evaluate how much any claimed violence represented gender inequality. But that certainly did not stop researchers from asserting exactly that. For me, any research group including WHO that tries to measure violence but questions only women and uses only women interviewers is clearly biased from the outset and findings cannot be trusted to be objective.
The WHO methodology may have been acceptable for exploratory probe studies to identify, for example, areas requiring scientifically rigorous clarification and quantification. But to disseminate the results as definite numerical measurements of actual violent victimization is dishonest and irresponsible. Yet that is exactly how the WHO study and its replications have been misused. Propaganda may be seen as useful in advancing a political cause, but effective solutions are only likely when a problem has been accurately and honestly elucidated.
Men will rightly be concerned that a large international organisation such as WHO, to which most countries contribute financially through the United Nations, would allow itself to be used as a dishonest propaganda machine in the feminist war against men.
Fanslow J, Robinson E. (2004). Violence against women in New Zealand: prevalence and health consequences. N Z Med J. 2004;117(1206). URL: http://www.nzma.org.nz/journal/117-1206/1173/
Fanslow J, Robinson E (2010). Help-seeking behaviors and reasons for help-seeking reported by a representative sample of women victims of intimate partner violence in New Zealand. Journal of Interpersonal Violence (May 2010), Vol 25:5, 929-951
WHO Multi-Country Study on Women’s Health and Domestic Violence: Initial results on prevalence, health outcomes and women’s responses (2005).